Every blast furnace campaign carries a quiet tension between two teams: the raw materials group that blends ore, sinter, and coke from whatever quality arrives that week, and the furnace operators who have to make that mixture behave predictably inside a vessel that never stops running. Burden distribution is the point where those two worlds collide, because the same charging program that worked perfectly last month can quietly push gas flow off-center the moment sinter strength or coke size shifts by a few percentage points. iFactory reads the incoming raw material data alongside real-time gas flow and temperature profiles to keep the ore-coke matrix correct charge after charge, and teams can see the model in action through Book a Demo.
Your Burden Program Was Tuned For Yesterday's Raw Materials
Ore grade, sinter strength, and coke size all drift week to week, yet most charging programs stay fixed until an engineer notices gas flow has already gone off-center. iFactory continuously recalculates the ore-coke matrix against live raw material quality so the burden profile adjusts before gas utilization drops, not after.
The Charging Program Is Fixed. The Raw Materials Are Not.
A blast furnace charging program specifies the ore-to-coke ratio, the batch weight, and the distributor chute angles for every layer, and it is typically set once by process engineering based on a target burden profile. The problem is that this program assumes the physical properties of the ore, sinter, and coke stay constant, when in reality every incoming shipment carries some variation in size distribution, reducibility, and mechanical strength. When a weaker batch of coke enters the furnace, it degrades faster on the way down the stack, generating more fines that block gas flow through the center. When sinter strength drops, it behaves differently under the burden's own weight, shifting the natural angle of repose and pushing more material toward the furnace wall than the program intended. None of these shifts trigger an alarm on their own, because each individual measurement can look within normal range while the combined effect on gas distribution compounds over dozens of charging cycles.
Peripheral Gas Flow Bias
A burden profile that runs slightly too permeable at the wall lets reducing gas escape upward along the periphery instead of passing through the ore column, lowering gas utilization and raising fuel rate without any single alarm firing.
Center Gas Flow Starvation
When coke concentration at the center falls below what the descending burden needs, central gas flow weakens, cohesive zone shape distorts, and the furnace becomes more prone to hanging and slipping events during the same shift.
Delayed Recognition of Drift
Because burden distribution effects accumulate gradually across many charges, operators often only notice a problem once top gas temperature spread has already widened well beyond the normal operating band for that furnace.
Manual Recalculation Is Too Slow
Recalculating chute angles and batch weights by hand for every raw material change is impractical during continuous operation, so most plants default to reacting only after a clear performance drop rather than adjusting proactively.
From Raw Material Data To Chute Angle In One Continuous Loop
iFactory ingests incoming raw material quality data as it arrives from the sinter plant, coke ovens, and pellet handling systems, then cross-references it against the furnace's live gas flow, top gas temperature distribution, and stockline profile from the burden probe. The model maintains a working estimate of how the current burden composition is behaving inside the stack, and it recommends adjusted chute angles, batch sequencing, and ore-to-coke ratios before the next charge is dropped, rather than waiting for a shift-end review to catch a drifting trend.
Stop Waiting For Gas Utilization To Tell You Something Went Wrong
iFactory recalculates your ore-coke matrix against live raw material quality so burden distribution stays correct charge after charge.
What Changes When The Burden Model Reacts In Real Time
The table below contrasts a traditional fixed charging program against an adaptive model that recalculates the ore-coke matrix against live raw material and gas flow data.
| Aspect | Fixed Charging Program | Adaptive AI Matrix |
|---|---|---|
| Response to raw material change | Manual review, often days later | Adjusted before the next charge |
| Gas utilization stability | Drifts with material quality | Held within target band continuously |
| Hanging and slip frequency | Reactive troubleshooting after events | Early warning from cohesive zone shape trend |
| Engineer workload | Manual recalculation for every shift | Recommendation review and approval only |
Furnaces with wider raw material variability, such as those blending multiple sinter or pellet suppliers, tend to see the largest gap between the two approaches because fixed programs cannot account for lot-to-lot differences.
What Feeds Into The Burden Distribution Model
The model's recommendations are only as good as the data feeding it, so iFactory pulls from several existing measurement points most furnaces already have installed, rather than requiring new instrumentation.
Burden Probe Stockline Profile
Radial stockline measurements show how evenly the burden surface is descending, which is the earliest physical sign of a distribution imbalance forming inside the stack.
Top Gas Temperature Array
A wide temperature spread across the top gas thermocouples indicates uneven gas flow, and the pattern of that spread tells the model which radial zone needs correction.
Raw Material Sizing And Strength
Coke drum index, sinter shatter strength, and ore size distribution from incoming lot testing set the baseline the model uses to predict how a batch will behave once charged.
Distributor Chute Position Log
Historical chute angle and rotation data lets the model correlate specific distributor settings with the gas flow outcome they actually produced for a given material mix.
We used to find out our burden was off-center only after top gas temperature spread had already widened for several shifts, and by then correcting it took days of careful adjustment. With iFactory reading our incoming sinter strength data against gas flow in real time, we now get a recommendation the same shift a raw material change starts to show an effect, well before it becomes a stability problem for the furnace.
What Furnaces Report After Adopting Adaptive Burden Control
The figures below reflect ranges reported by furnaces after adopting a burden distribution model that recalculates against live raw material and gas flow data, and they vary by furnace size, raw material variability, and existing instrumentation quality.
Common Mistakes When Introducing Adaptive Burden Control
Furnaces that try to introduce adaptive burden control without preparing the surrounding process tend to run into a few recurring issues.
Ignoring Raw Material Data Latency
If incoming lot testing results arrive hours after material is already staged for charging, the model's recommendations lag behind the material actually entering the furnace, undermining the whole point of adaptive control.
Overriding Recommendations Without Feedback
When operators override a recommendation without logging why, the model loses the chance to learn from that judgment call, slowing down how quickly it adapts to that furnace's specific behavior.
Skipping Distributor Calibration Checks
A distributor whose actual chute angle drifts from its commanded position over time will make even a correct recommendation land in the wrong place on the burden surface.
Treating It As Fully Automatic Immediately
Furnaces that see the fastest, safest adoption keep an experienced operator in the review loop for the first several months before trusting recommendations without confirmation.
Frequently Asked Questions
Q: Does this require new sensors or instrumentation on the furnace?
In most cases, no. The model is built to work with instrumentation furnaces already have installed, including the burden probe, top gas temperature array, and existing raw material lot testing from the sinter plant and coke ovens. Where a furnace is missing a specific measurement point, such as a full radial temperature array, the model can still operate on a reduced dataset while flagging where additional instrumentation would improve recommendation accuracy. Reach out through Support Contact to review what your furnace already has in place.
Q: How does the model handle a sudden raw material supplier change?
A new supplier's material typically carries different sizing and strength characteristics than what the model has previously seen, so recommendations during the transition period are generated with wider caution margins until enough charges confirm how the new material actually behaves inside that specific furnace. This is one of the situations where keeping an operator in the review loop matters most, since early judgment calls during a supplier transition help the model calibrate faster than it would from charge outcomes alone.
Q: Can this reduce hanging and slip events, or only improve gas utilization?
Both outcomes are connected, because hanging and slip events are frequently downstream consequences of a burden distribution that has drifted enough to disrupt cohesive zone shape. By catching the earlier gas utilization drift before it compounds into a full instability event, the model addresses the root cause rather than only the symptom, which is why furnaces typically report improvement in both metrics together rather than one without the other.
Q: How long does it take before the model's recommendations are reliable for our specific furnace?
Every furnace has a distinct relationship between chute settings, raw material characteristics, and resulting gas flow, so the model needs a calibration period against that specific furnace's charging history before recommendations reach full confidence. Most furnaces see meaningfully useful recommendations within the first few weeks, with accuracy continuing to improve as more charge outcomes are logged. A Book a Demo session can walk through a realistic calibration timeline for your furnace's raw material profile.
Q: Does adaptive burden control replace the process engineer's role in setting the charging program?
No, it changes what the process engineer spends time on rather than removing their role. Instead of manually recalculating chute angles for every raw material shift, the engineer reviews and approves model-generated recommendations, focusing judgment on unusual situations like a new supplier or an equipment change rather than routine recalculation for known material variation.
Keep Your Ore-Coke Matrix Correct, Charge After Charge
iFactory reads incoming raw material quality against live gas flow so burden distribution adjusts before performance drifts, not after.
